Heriot-Watt University
ROS: The Research Output Service. Heriot-Watt University EdinburghNot a member yet
4689 research outputs found
Sort by
Enhancing the understanding of cost overrun drivers in highway projects in Nigeria through system dynamics modelling
Get PDFHighway projects in developing countries (and specifically in Nigeria) are beset with
numerous implementation challenges, including delivery within the contracted budget.
This particular problem can be attributed to poor understanding of the contextual
interactions and overall dynamics of the factors which contribute to cost overrun. To
overcome this challenge, efforts have been made to understand the drivers of cost overrun,
but to date there has been insufficient focus on the intricate interactions and dynamics of
the driving factors. Therefore, this study set out to assess how system dynamics approach
could improve our understanding of the driving factors of cost overruns in highway
projects in Nigeria, by developing a computer-based system dynamics model that
incorporates the intricate and contextual dynamic processes of the drivers. Using a case
study research strategy, a four-stage interview process was adopted. Firstly, 16 interviews
with relevant stakeholders were conducted and analysed using a formulated data
compatible coding framework to inform model conceptualisation. Secondly, 9 interviews
were conducted to validate the developed conceptual model. Thirdly, 9 interviews were
conducted to elicit the values of the parameters used in the formal modelling process.
Lastly, 6 interviews were conducted to validate the simulation model by presenting the
simulation results and the accompanying algorithms and documentation. The findings
revealed that, amongst the numerous driving factors, the most significant were: delay in
progress of work, political instability, adverse weather, social issues, modification of
project scope, and delay in payment to contractors. Additionally, the multiple drivers
were shown to exhibit causal relationships, which formed the basis for the development
of a conceptual model, supported by identified causal relations from literature sources
and a model evaluation process with the stakeholders. The results from the conceptual
model indicated that the feedback structure of the contextual system is governed by seven
feedback loops (3 reinforcing (positive) and 4 balancing (negative)), demonstrating the
complexity and dynamics of the contextual system. Accordingly, key parts of the
conceptual model (Loops R3 and B4) which described the dynamic interactions inherent
in a typical highway infrastructure development in Nigeria were converted, expanded and
numerically structured in terms of stock and flow diagram, and simulated in a 240-month
time period. The simulation results suggested that the model is useful for its purpose and
is consistent with the system’s observed contextual reality. As well as the business as
usual (BAU) scenario, which assumed the status quo in contextual behavioural trends,
three alternative scenarios were designed and simulated in order to assess their impact on
improving the cost performance of highway projects in Nigeria. The results showed that
scenario four (combination of all policies) would provide the maximum cost performance
benefit, by ensuring that the delivery period of key economic projects improved by 15%.
The evaluation results indicated that the processes and their outputs are vital instruments
for enhancing and promoting a better understanding of the intricate and contextual
dynamics of cost overrun drivers in highway projects. This new model will be of use to
decision makers in the highway sector of the Nigerian construction industry, facilitating
informed and pragmatic decisions regarding policies that will ensure cost effective
delivery of highway projects
Advanced mechanical and radiofrequency design of reflectarray antennas for space applications
Get PDFThe analysis, optimization, and design of advanced reflectarray (RA) antennas for space
applications are investigated in this work. The study is focused on developing methodologies
and tools able to be integrated into the industrial framework of this thesis.
RA antennas are a relatively emerging technology in the space telecommunication framework, benefiting from the advantages of the conventional reflectors and array antennas for
high gain applications. This work surveys the coupled mechanical and radiofrequency (RF)
design aspects of RA antennas. The mechanical design of large RA antennas shows some
limitations concerning the survivability in the harsh space environment. Indeed, strong
thermoelastic deformations of the RA composite structures can remarkably affect the antenna’s electrical performances. This work surveys these aspects by proposing reliable and
cost-effective structural solutions, which are consequently taken into account in the RA
electrical design process. The electrical analysis and design process benefits from internal
industrial tolls that are efficiently integrated with original tools developed in this thesis
framework.
The electrical design process lies in some hypotheses that allow an efficient design and op timization of the RA antenna layout, such as the local periodicity assumption. The local
periodicity hypothesis is well respected in the RA design process investigated in this work
by employing a particular unit cell, named Phoenix cell. The Phoenix cell geometries are
efficiently handled in parametrized lookup tables to prevent any sharp geometrical transition on the RA layout. This work proposes and compares different design and optimization
approaches in which the unit cells definition, and the subsequent lookup table construction,
are efficiently integrated. The classical phase-only synthesis is surveyed and implemented.
It represents the departing point for the direct layout optimization, aiming to improve
the RA performances by naturally preventing any sharp geometrical transition on the RA
layout. The optimization is conducted through a continuous and regular modulation of
the RA cells geometries distribution by targeting prescribed performances, either on the
aperture or the far-field.
The coupled RF/mechanical design is applied to a concrete case of a large deployable
RA for telecommunication spacecraft. Moreover, the electrical analysis and design of a
small satellite RA antenna for payload data handling and transmission are investigated
and implemented
Dynamics of dissolved organic matter composition in Scottish rivers and headwater streams – resolving environmental and biogeochemical process interactions
Get PDFDissolved organic matter (DOM) has a wide range of chemical structures that give it a
multifunctional role in the natural environment. Although the role of DOM in aquatic
ecosystems has been the focus of previous work, a comprehensive understanding of the
compositional behaviour of DOM under different environmental processes is still
incomplete. New field-based geochemistry data is presented from a two-year study
(03/2017- 03/2019) in Scottish headwaters and a 9-month study in large Scottish rivers.
This research shows that the DOM mobilisation follows seasonality with enhanced
exports of DOM during winter months compared to the summer. At a larger spatial scale,
the seasonal trend is overprinted by the catchments soil type. Size-Exclusion
Chromatography combined with high-resolution time series of DOM variables reveal that
precipitation events preferentially mobilise humics from the surrounding soils, while
humics concentration decline during low flow conditions. Furthermore, the data show
that non-UV absorbing (“invisible”) low molecular weight (LMW) neutrals (iDOM)
contribute up to 50 % to the total DOM pool in headwaters, especially during low flow
conditions, and on average 13 % to the DOM in larger river systems. The source of iDOM
was found to be the topsoil of peatland and peaty podzols. Consequently, more labile OM
can be leached from soils into the aquatic environment in the future through disturbed
soils promoting instream microbial growth and act as a nutrient source for aquatic plants
Evolving marketing strategies for Swiss ICT SMEs : a marketing strategy canvas in the light of digital transformation
Get PDFIt is crucial that small and medium sized businesses in the ICT sector adopt strategies to
enable them to exploit, capitalise upon, and respond to the digital transformation. SMEs
will need to undergo profound and continuous changes, developing more strategic
focused marketing thinking and business models to enable or find new markets while
attaining and retaining competitive advantage.
SMEs have to fulfil the same marketing tasks as large companies, but under conditions
characterised by smaller sales volumes and small market niches. The major difference for
SMEs relates to the lower availability of resources, especially financial and human
resources. This constraint makes it even more important to define an appropriate
marketing strategy. SMEs need to adopt marketing strategies that build on their main
advantages such as close proximity to customers, operational flexibility, flat hierarchies,
short decision-making paths and rapid reaction times. The researcher addresses the
research question: ‘How should Swiss SMEs in the ICT sector define their marketing
strategy to benefit from digitalisation?’ by adopting a constructivist philosophy utilising
a qualitative approach. Although this has been generally researched for large enterprises,
SMEs with their specific characteristics should also be accounted for. A conceptual
framework was formulated to inform the development of a marketing strategy for SMEs
in the digital context. The research employed a mono methodological research design
involving 14 case studies, conducted across businesses that have successfully transformed
for the digital age. One marketing expert and two marketing agencies were additionally
considered to strengthen the credibility and plausibility of the case studies.
The findings indicate that changes in the market situation and the rapid evolution of
various modern media channels require new levels of personalisation and targeted
campaigns that drive engagement and brand loyalty. The study contributes to the existing
theoretical body by arguing that the current view of marketing strategy in the digital era
needs to be reformed to emphasise on the customer journey and to add data-driven
elements to the mix using key digital marketing metrics. Further, the study suggests that
ICT SMEs can improve customer experience and gain competitive advantage by
designing new customer touchpoints through content. The study makes a practical
contribution by proposing a marketing strategy canvas with eleven dimensions as a tool
to help SMEs strategise digitisation and marketing strategy
Advanced direct metal 3D printed passive components for wireless communications and satellite applications
Get PDFThis thesis presents the design of advanced microwave passive filters, antennas, and
antenna arrays using direct metal 3D printing technology. These work all incorporate the
printing technology into the RF component design process, demonstrating the potential
possibilities of direct metal 3D printing in the investigation and fabrication of passive
microwave components with irregular shapes but attractive features.
This thesis's works involved an extensive frequency range that starts with investigating S-band filters and then extends to C-band and Ku-band filters and antennas design. It is well
known that in S- and C- band radio frequency (RF) applications that miniaturization is a
critical factor for RF devices besides high performances. For this reason, the first project
in this thesis proposed a novel compact waveguide loaded air slots resonator for designing
inline bandpass filters. As a result, the designed filters not only have a smaller size than
coaxial ones but also have controllable transmission zeros with inline structures. Since the
air slots resonator is loaded inside the cavity, it is difficult to fabricate by conventional
methods, but accessible by 3D printing technique with appropriate self-support structures.
The fabrication quality was reflected by the mechanical and RF property measurements,
which first demonstrated the advantage of using 3D printing technique to fabricate
components with complex structures.
The second project presents a compact high-Q fan-shaped folded waveguide resonator,
which is applied to successfully design one C-band filter and filtering antenna. High
performance RF properties and easy-to-print structures are always considered together.
Accordingly, this work proposed and validated novel slots cross negative coupling
topology of the filter and novel filtering antenna theory. Also, each of the designed
components has better self-supported structures that can be printed with only two pieces,
which highly reduced assembly processes and errors. Furthermore, the RF properties from
measurement results further demonstrated that the reliability of the metal 3D printing
technology for C-band RF applications.
The concepts of the third project are extended from the second project but replaces the
folded waveguide resonator with a metal strong coupling resonator (MSCR). The MSCR
allows for even further compact dimensions while maintaining a high Q value of over
1000. It also allows producing mixed electrical-magnetic coupling by the curving coupling
metal pairs intentionally. Except for the desired RF properties, the designed filter based on
the MSCR can be printed as a whole even with complex inner circuits structures.
Furthermore, the MSCR was integrated with the helical antenna using the proposed theory
presented in the second project. Although the helical antenna belongs to the electrical-small antenna, the designed filtering antenna still has a high transmission efficiency of
more than 95% and a 6 dBi realized gain concerning its less than quarter-wavelength. In
addition, the filtering antenna has five helical radiation elements and one filter prototype
but was printed with only three pieces, which showed the advantages of the direct metal
3D printing technology again.
The fourth and the last project introduces a Ku-band slots antenna array application based
on the sine corrugated waveguide resonator. Similar to previous projects, advanced RF
performances were pursued in this project, in addition to demonstrating the use of 3D
printing technology to fabricate compact and specific structures. The designed antenna
array achieved a higher gain, wider band, and more simple feeding networks. The mode
analysis method based on the EM software CST was applied to guide the design since no
related formulas were available. The designed model was printed with two pieces and was
measured thoroughly. The measured surface roughness, in-band responses, and radiation
patterns showed promising results for the sine corrugated waveguide and 3D printing
technology in satellite applications.
In general, this thesis researched and proved the reliability and advantages of direct metal
3D printing technology in designing and fabricating advanced microwave passive
components below the Ku-band. It should be mentioned that the designed passive
components in this thesis can be easily re-designed/re-configured and applied on the 5G
wireless base station and satellite communication systems
What shapes cross-border merger and acquisition negotiations in the automotive industry?
Get PDFThe research evaluated the impact of contextual, structural, and behavioural factors in
shaping cross-border merger and acquisition (CBMA) negotiations between automobile
manufacturers. Recent years have seen an increase in CBMA activity in the automotive
industry, advanced by the necessity to share investments in alternative power sources for
engines and realise economies of scale and scope. The significance of the topic is reflected
by the essential role played by the automotive industry in the global economy. According
to Fortune (2020), the combined revenue of the top 10 automakers exceeded 1.70 trillion
in 2019. PricewaterhouseCoopers (2020) reported that Global Automotive M&A activity
accounted for 100 billion and approximately 800 deals in 2018, and over $77 billion and
around 850 deals in 2019. Despite the substantial deal value and volume, research has
repeatedly determined that over 70 per cent of CBMAs fail to deliver the promised results
due to the ineffective management of the negotiation process. Moreover, while the
different stages of the M&A process have been extensively investigated, research on the
M&A negotiation phase has been limited, and very few studies have attempted to
incorporate contextual, structural, and behavioural factors in analysing inherently
complex CBMA negotiations.
The study followed a pragmatist standpoint and adopted a sequential mixed-method
design integrating the macro-strategic and micro-behavioural levels of analysis. The first
phase based on a qualitative small-N focused comparative analysis case study on
identifying the type of precipitant originating turning points in CBMA negotiations
between automobile manufacturers. The second quantitative phase entailed a factorial
experimental design and questionnaires to evaluate motivational and relational factors'
role in shaping the negotiators' response to the previously identified precipitants. The
simulations extensively conformed to a real case, and the sample of experimental
participants consisted of executives with at least seven years of negotiation experience.
The findings indicate that negotiation outcomes are significantly influenced by elements
internal to the negotiation process, with contextual factors (including culture) exhibiting
only a marginal influence. The conclusions also highlight the critical role of coalition-building in shaping the negotiation process. The results supplement current literature and
provide a roadmap for managers to better prepare, identifying the three crucial
behavioural factors that shape negotiators' response to precipitants and significantly
influence the outcome of CBMA negotiations between automobile manufacturers: the
seller's motivation and power perception and the buyer's affective trust
Actor-critic reinforcement learning algorithms for yaw control of an Autonomous Underwater Vehicle
Get PDFAn Autonomous Underwater Vehicle (AUV) poses unique challenges that must be solved
in order to achieve persistent autonomy. The requirement of persistent autonomy entails
that a control solution must be capable of controlling a vehicle that is operating in an
environment with complex non-linear dynamics and adapt to changes in those dynamics.
In essence, artificial intelligence is required so that the vehicle can learn from its
experience operating in the domain.
In this thesis, reinforcement learning is the chosen machine learning mechanism. This
learning paradigm is investigated by applying multiple actor-critic temporal difference
learning algorithms to the yaw degree-of-freedom of a simulated model and the physical
hardware of the Nessie VII AUV in a closed-loop feedback control problem. Additionally,
results are also presented for path planning and path optimisation problems. These control
problems are solved by modelling the AUV’s interaction with its environment as an
optimal decision-making problem using a Markov Decision Process (MDP).
Two novel actor-critic temporal difference learning algorithms called Linear True
Online Continuous Learning Automation (Linear TOCLA) and Non-linear True Online
Continuous Learning Automation (Non-linear TOCLA) are also presented and serve as
new contributions to the reinforcement learning research community. These algorithms
have been applied to the real Nessie vehicle and its simulated model. The proposed
algorithms hold theoretical and practical advantages over previous state-of-the-art
temporal difference learning algorithms. A new genetic algorithm is also presented
and developed specifically for the optimisation of the continuous-valued reinforcement
learning algorithms’. This genetic algorithm is used to find the optimal hyperparameters
for four actor-critic algorithms in the well-known continuous-valued mountain car
reinforcement learning benchmark problem.
The results of this benchmark show that the Non-linear TOCLA algorithm achieves a
similar performance to the state-of-the-art forward actor-critic algorithm it extends while
significantly reducing the sensitivity of the hyperparameter selection. This reduction in
hyperparameter sensitivity is shown using the distribution of optimal hyperparameters
from ten separate optimisation runs. The actor learning rate of the forward actor-critic
algorithm had a standard deviation of 0.00088, while the Non-linear TOCLA algorithm
demonstrated a standard deviation of 0.00186. An even greater improvement is observed
in the multi-step target weight, λ, which increased from a standard deviation of 0.036 for
the forward actor-critic to 0.266 for the Non-linear TOCLA algorithm. All of the sourcecode used to generate the results in this thesis has been made available as open-source
software.ARchaeological RObot systems for the Worlds Seas (ARROWS) EU FP7 project under grant agreement ID
30872
Exposure to environmental and occupational particulate air pollution and their association with diabetes and dementia : a cross-sectional epidemiological analysis in UK Biobank
Get PDFWork-related illness is an important societal problem and the number of new cases of a variety
of associated diseases is increasing annually around the world. Despite this there are almost
certainly a number of health conditions caused by work that are still unrecognised. There is
evidence that environmental particulate air pollution causes diabetes and dementia, but at this
time it is uncertain whether similar occupational exposures can also cause these diseases. In this
research possible associations between workplace aerosol exposure and diabetes and dementia
are investigated.There are many occupations where workers are exposed to airborne particles.
Some exposures in the workplace are very similar to environmental pollution, mainly where
there is potential exposure to combustion particle aerosol, such as diesel soot or fume. In
workplaces, ultrafine particles are also found in metal and polymer fumes, both of which can
induce acute inflammatory responses in the lung. Occupations exposed to these kinds of
pollutant include construction workers, tunnel workers, miners, metal workers, farmers using
diesel equipment, wood burners, road-patrol officers, people who work in parking lots or
automobile factories and generally workers in open areas in cities.
A systematic review to spot the gaps for workplace exposures (Study 1) and then an
epidemiological analysis was carried out using the UK Biobank resource to investigate the
association between occupational and environmental exposure to fine airborne particulate matter
in relation to both diabetes (type-2 diabetes) and neurodegenerative diseases (dementia) (Study
3). Existing Biobank Job-Exposure Matrix (JEM) estimates were refined in order to be relevant
for our needs (Study 2). It was hypothesised that the investigation of the effects of environmental
air pollution exposure in the UK Biobank cohort would confirm the results from other
epidemiological studies investigating the adverse health effects of environmental pollutants, and
that investigation of analogous occupational exposures in the same cohort would provide insight
as to whether similar risks arise in the workplace.
The studies in this thesis have addressed gaps in knowledge related to environmental and
occupational particulate exposure and type-2 diabetes mellitus (T2DM) and dementia. Obtained
data suggests an association between environmental particulate air pollution exposure and
T2DM. We did not find a significant association between environmental particulate air pollution
exposure and dementia, and this could be because of the relatively low dementia prevalence in
the cohort. There is no strong evidence for an association between occupational exposure to
particulates that are similar to environmental air pollution and T2DM or dementia
Evaluating associations between urban greenness and health outcomes
Get PDFThe United Nations reported that 55 percent of the world’s population lived in urban
areas in 2018, and that such proportion is expected to increase to 68 percent by 2050. Since the percentage of people living in the urban environment continues to rise, there is
an urgent need to better understand the relationship between green space and a wide
range of health indicators with a specific focus on environmental and health inequality. The rapid development of urban and rural areas has brought not only environmental
inequality, but also spatial injustice, which has posed a continuing threat to human
health and wellbeing. In this thesis, the author provides a framework where potential
pathways by which urban greenness contributes to health benefits are generated into
three dimensions of measuring the green space: satisfaction with green spaces, proximity to green spaces, and frequency of visiting green spaces. The thesis is based in
Beijing, China. The first dimension of this thesis aims to examine the association between satisfaction
with two types of green space and residents’ self-rated health by comparing
neighbourhood green space (NGS) and community green space (CGS) across spatial
dimensions, which are urban and suburban dimensions. 4291 respondents were derived
from a large-scale individual survey of inhabitants of Beijing city in 2013 and provide
the basis for the analysis. Multilevel ordered logistic regression analysis is used to
examine the associations between residents’ satisfaction with the two types of green
spaces and residents’ self-rated health. The author finds that residents who are more
satisfied with NGS and CGS have higher odds of reporting good self-rated health
outcomes. Such effects are more pronounced for residents living close to NGS and tend
to decline non-linearly over space. Additional results quantify the differentiated effects
on self-rated health between urban and suburban residents. The findings of this
dimension suggest that residents’ satisfaction with different types of green space on
health benefits should be taken into account in the land use design of green space
preservation and development policies. The second dimension of this thesis aims to explore the association between residential
proximity to public green space (parks), the semi-public green space (Olympic Park), private green space (golf courses) and residents’ self-rated health. It further explores
whether the perceived spillover effects of public parks on health benefits may be influenced by access to golf courses and the Olympic Park. 4762 respondents from a
large-scale individual survey of inhabitants of Beijing city in 2013 provides the basis for
the analysis of this subsection. Similar to the first dimension, multilevel ordered logistic
regression analysis is also used to examine the associations between different types of
green spaces and residents’ self-rated health. The author finds that residential proximity
to the Olympic Park and golf courses have higher odds to report being good self-rated
health outcomes compared to those that do not. This association tends to decline with
distance away from the Olympic Park and golf courses in a non-linear manner. A
comparison of model specifications for residential experience, housing, and income
characteristics suggests that different segments of residents have different self-rated
health correlates of proximity to different green spaces. Complementary effects between
different types of green space and amenities highlight the importance of considering the
needs of proximity to green space for marginalised and disadvantaged social groups, such as migrants and low-income people. This should be considered when policy
makers are designing land use policies. The last dimension of this thesis aims to explore the overall relationship between
frequency of using green spaces, perceptions of neighbourhood environment quality and
mental health outcomes, while taking into account the influences of demographic
characteristics and lifestyle status. The author performed a comparative case study in
four different communities in Beijing city that vary in the accessibility to green space
measured with service radii. It aims to examine the direct effects of the frequency of
visiting different green spaces and perceptions of green space quality on residents’ mental health outcomes. Structural equation modelling (SEM) was applied to explore
such effects. Data was collected by combining the face-to-face questionnaire with a
web-based questionnaire due to the outbreak of COVID-19. Overall, 115, 115, 111 and
100 responses were obtained respectively from four different communities. As a result, the author finds a positive association between residents’ frequency of visiting green
space and better mental health. Additional results suggest that the positive effects of
residents’ frequency of visiting green space on mental health outcomes decay by green
space measured with different service radii. Findings shed light on providing a potential
pathway for stakeholders to maintain the balance of preservation of urban green spaces
to mitigate the influence of rapid urbanization on people’s health outcomes
Embedded intelligence for self-test of a MEMS microphone
Get PDFThe reliability of micro-electro-mechanical-systems (MEMS) microphones have
increased over the past decades. While other MEMS devices, such as accelerometers and
radio frequency (RF) MEMS switches, already have built-in-self-test (BIST) and built-in-self-repair (BISR) solutions, there is no evidence of such research in the field of MEMS
microphones. This thesis therefore presents in-field suitable BIST methods that allow the
extraction of specific information regarding device health as a way to further increase the
reliability of single diaphragm capacitive MEMS microphones.
Following the first literature review since 2006 on BIST for MEMS, six appropriate
methods are proposed, although none of them offer complete fault coverage.
Consequently, two BIST methods are proposed in this study: a system-level “loud BIST”
solution requiring a dedicated test signal from a sound source in the system, and a “silent
BIST” solution utilising ambient noise as a sound source, dismissing the need for a
dedicated speaker or a test signal.
To assess the feasibility of such BIST solutions, a machine learning-based four-layer
framework is developed, representing a methodology for the creation BIST and BISR
solutions with: lumped element modelling to assess the behaviour of the microphone,
failure mode simulation to characterise the behaviour of the defects affecting the
frequency response of the microphone, machine learning and global minimum finding based failure mode discrimination to identify the failure modes affecting the microphone
solely based on the frequency response of the DUT and recovery to restore the
functionality of the microphone after a non-catastrophic fault.
The experimental validation of the framework is achieved through artificial non-catastrophic induction of failure modes that include diaphragm stiction, lid attach holes,
cracked membrane, broken vents and blocked porthole, with the frequency response of
the microphones captured before and after defects. Measurements were carried out both
on laboratory stage and real-life acoustic setups built for the demonstration of the
feasibility of in-field deployment of the proposed BIST methods. Through the obtained
results, failure mode discrimination is demonstrated with and without a dedicated test
signal for both laboratory and real-life conditions.Engineering and Physical Sciences Research Council (EPSRC) fundin
Access Repository Dashboard
Do you manage ROS: The Research Output Service. Heriot-Watt University Edinburgh ? Access insider analytics, issue reports and manage access to outputs from your repository in the CORE Repository Dashboard!